AI Security Architect

Quarks Technosoft Pvt. Ltd.; Pune District; On-site; INR 4,000,000 - 7,000,000; Full time

Job summary

Quarks Technosoft Pvt. Ltd. is seeking a senior cybersecurity leader to shape security architecture for AI-driven enterprise systems. The role emphasizes hands-on design reviews, threat modeling, and governance across cloud, data, and agent-based platforms.

The ideal candidate combines deep security design with the ability to brief executives and manage security posture under commercial pressure, especially in AI-enabled environments.

Qualifications

• Bachelor’s degree or equivalent professional experience.
• Minimum 10 years in cybersecurity, with at least 4 years in security architecture or equivalent capacity.
• Hands-on involvement in securing production AI, generative AI or agentic systems beyond advisory exposure.
• Deep experience in cloud and enterprise security architecture, with Azure and Microsoft security tools preferred.
• Working technical understanding of LLM architectures, embeddings, vector stores, RAG pipelines, and agent frameworks.
• Ability to produce threat models and control mappings without external support.
• Experience designing or governing agentic systems that act on enterprise data, including autonomy tiers and human-in-the-loop controls.
• Proven track record of authoring security standards and architecture documentation for enterprise scale.
• Identity and data security grounding including non-human identities and data loss exposure through AI interfaces.
• Hands-on prototyping with current AI development tooling and self-hosted models.

Skills

Security architecture

AI security

Cloud security

Threat modeling

Executive briefing

Python scripting

Education

Bachelor’s degree or equivalent experience

Tools

Azure

Microsoft Defender / SIEM

XDR platforms

Job description

Essential qualifications and experience:

• Bachelor’sdegree in any discipline, or equivalent professional experience. Technicaldisciplines are relevant but not required; professional certification anddemonstrated practice carry equal weight at this level.
• Minimum10 years in cybersecurity, of which at least 4 years in a security architectureor equivalent technical design authority capacity.
• Minimum2 years of hands-on involvement in securing production AI, generative AI oragentic systems — not solely policy or advisory exposure.
• Demonstrabledepth in cloud and enterprise security architecture; Microsoft Azure and theMicrosoft security platform are strongly preferred given the enterprise estate,including XDR, SIEM, cloud posture and the AI and agent platform serviceslayered on them.
• Workingtechnical understanding of LLM architectures, embeddings, vector stores, RAGpipelines, fine-tuning, agent frameworks, multi-agent orchestration and theMCP-style tool integration model — sufficient to challenge an engineeringteam’s design on its merits.
• Practicalcommand of AI threat taxonomies and the ability to produce threat models andcontrol mappings without external support.
• Evidenceof designing or governing agentic systems that act on enterprise data orsystems, including autonomy tiers, trust boundaries and human-in-the-loopcontrols.
• Provenrecord of authoring security standards, patterns and architecture documentationadopted at enterprise scale.
• Identityand data security grounding: non-human, agent and workload identity, privilegedaccess, data classification, DLP and data loss exposure through AI interfaces.
• Hands-onprototyping capability using current AI development tooling, includingagent-building platforms, AIassisted coding environments and locally hosted orself-managed models for evaluation work. The role is expected to build toassess.
• Abilityto write and read Python to the level required to validate a control, review atest harness or prototype an assessment.
• DemonstratedC-level advisory experience: maturity assessment, transformation roadmapdevelopment and executive briefing, with examples of decisions influenced.
• Crediblewith both engineers and executives, and able to hold a security position undercommercial pressure.